The ignition process is a critical consideration for combustion devices. External energy transfer to the combustor is required for ignition in common combustion systems. There are many ways to deposit energy into the flow but a standard method is a spark discharge because it is simple, compact, and reliable. Sparks can be categorized as either inductive or capacitive sparks that use a coil or an electrical resonance circuit with capacitor, respectively, to amplify the voltage. The creation of a successful ignition event depends on the spark energy deposited into the flow, the initial composition, pressure, temperature, turbulence level of flow etc. The deposited energy by the spark into the flow is critical for estimation of initial energy available for ignition of the mixture. Therefore, the electrical characteristics of the sparks were investigated under various flow conditions. Then measurements of deposited energy into the flow were conducted using a very accurate experimental procedure that was developed in this research. The results showed considerable electric energy losses to the electrodes for the relatively long, inductive sparks. However, the short, capacitive spark deposits electric energy into the flow with minimal loss (above 90% deposition efficiency). In addition, the characteristics of inductive spark are affected by flow velocity and by the existence of a flame. However, variations in the flow conditions do not affect the characteristics of the capacitive spark such as voltage-current time trace and energy deposition efficiency. Two ignition systems using above mentioned two spark types were developed. First, the capacitive spark energy was directly deposited into the premixed flow. Most researchers have not concentrated on the early initiation process but on the flame growth. Therefore, the generated kernel formed by the energy deposition was observed and characterized using optical methods, immediately following the spark. In addition, the mixing effect for this ignition kernel with surrounding gas was simulated using a numerical method. Based on the time trace of the OH* chemiluminescence, the reaction starts with the discharge and it is continuous until combustion begins. This means that in the presence of a high density spark in premixed flow, there exists no traditional delay as defined by other researchers for auto ignition. A simple Radical Jet Generator (RJG) was developed that is able to ignite and stabilize a flame in a high-speed flow. The inductive spark initiates the combustion in the RJG chamber. The RJG then injects the partially-burned products carrying large amounts of heat and radicals into a rapidly moving flammable main stream. Then it ignites and stabilizes a flame. The RJG requires low levels of electrical power as long as the flow velocity is relatively low since most of the radicals are produced by the incomplete combustion in its chamber. The importance of radicals was analyzed by RJG experiments and numerical methods. The reaction zone for RJG using a rich mixture was located both inside and outside of the RJG chamber. Therefore, the RJG using a rich mixture performed better in the ignition and stabilization of combustion in the main flow. According to an analysis using the CHEMKIM simulation software combined with the San Diego chemical mechanism, the RJG jet resulting from a rich mixture contains more radicals and intermediates than that produced by a lean mixture for the same sensible enthalpy. In addition, the burned gas contains less radicals and intermediates than the partially burned gas. If the RJG is operating with a high speed main flow, the flow rate through the RJG chamber must be increased to allow the radical jet to penetrate well into the rapid flow due to their higher injection velocity. Unfortunately, this leads to unsteady combustion in the RJG, which results in the pulsation of the radical jet. This reduces the number of radicals injected into the main flow. To investigate this operating condition, special attention was focused on four possible factors: unburned reactant pockets caused by motion of the spark channel, spark frequency, flame propagation speed and ignition delay. It was shown that the unsteadiness is affected by the flame speed and ignition delay because the frequency of pulsation in the chamber is highly dependent on the equivalence ratio. In addition, the interaction between the RJG operation and the combustion dynamics in the main combustor was documented. The acoustic pressure oscillations in the main combustor were suppressed when the RJG jet was turned on because the reaction region is relocated by the operation of the RJG.